Peloruside A, a Novel Antimitotic Agent with Paclitaxel-like Microtubule- stabilizing Activity¹ Free

In the search for anticancer drugs, compounds from natural sources, such as paclitaxel, extracted from the bark of the Pacific yew tree, have displayed clinically useful anticancer activity. Peloruside A (peloruside), a novel cytotoxic compound, was recently isolated from the marine sponge Mycale hentscheli (1). This same species of sponge also produces two other potentially useful secondary metabolites, mycalamide A and pateamine, with very different structures and bioactivity profiles as compared with peloruside (2). Peloruside is a potent cytotoxin at nanomolar concentrations, inducing biochemical changes consistent with apoptosis in a number of cultured mammalian cell lines (3). The strategy of using tubulin as a target for cancer chemotherapy is based on the increased growth and division of cancer cells and the fact that drugs that interfere with mitosis such as the Vinca alkaloids that shift the equilibrium to the depolymerized form of tubulin have proven effective in the treatment of cancer. Paclitaxel (Taxol®) and taxotere (Docetaxel) target tubulin but, unlike the Vinca alkaloids and colchicine, shift the equilibrium to the polymerized form, thus stabilizing microtubules. Both are currently used therapeutically for the treatment of solid tumors of the breast, ovary, and lung (4, 5). Microtubule-stabilizing compounds can be divided into three groups: (a) the terpenoids (including the taxanes, paclitaxel, and taxotere) isolated from yew trees (4), eleutherobin/sarcodictyin isolated from marine corals (6), and the bacterial metabolite WS9885B (5); (b) the macrolides, including epothilones, isolated from the bacterium Sorangium cellulosum (7, 8) and laulimalides isolated from the marine sponge Cacospongia mycofijiensis (9); and (c) the polyhydroxylated alkatetraene lactones, including discodermolide, isolated from a Caribbean sponge (10, 11). Peloruside is a macrolide similar to epothilone containing a 16-membered ring, whereas the laulimalides have a 20-membered ring. The complex chemical syntheses required to produce clinically useful amounts of the above-mentioned drugs have limited their development as anticancer agents, although both epothilone and the more complex paclitaxel and taxotere have now been synthesized in sufficient amounts for clinical use. In addition, paclitaxel is lipophilic and thus has low aqueous solubility. For clinical use, it must be dissolved in polyoxyethylated castor oil (Cremophor EL), a vehicle that contributes to paclitaxel’s undesirable side effects, in particular its hypersensitivity reactions (7). Paclitaxel’s hydrophobicity also promotes the acquisition of the multidrug resistance phenotype through expression of P-gp⁵ (12). P-gp is responsible for the efflux of a broad range of organic solutes from the cell, and paclitaxel is just one of these. In addition to overexpression of P-gp, some cells become resistant as a result of mutation of the paclitaxel binding site on β-tubulin (13). The search is therefore on to find other microtubule stabilizers that have similar antimitotic activity to paclitaxel but that lack the interaction with P-gp or bind to unique sites on the tubulin polymer. Epothilones, laulimalides, and discodermolides have shown promise in this area, displaying less loss of toxicity to certain P-gp-expressing cells than paclitaxel (5, 7, 8, 9, 11), although still being transported to some extent by P-gp. At least three of the known microtubule-stabilizing drugs, the epothilones (7, 8), discodermolide (11), and the eleutherobins (6), compete with [³H]paclitaxel for its binding site on β-tubulin; however, epothilone and discodermolide show different sensitivities to particular β-tubulin mutations despite binding to a similar site. The paclitaxel binding site of β-tubulin is available at 3.5 Å resolution (14), facilitating drug modeling approaches. A common pharmacophore has been partially described, but further structure/function studies are needed (4, 5, 13, 15, 16). Recent in vivo tests on tumor formation in nude mice have shown promise for desoxyepothilone analogues, specifically Z-12,13-desoxyepothilone B and its more water-soluble analogue, Z-12,13-desoxyepothilone F (5, 17). Both analogues are less toxic and have an improved therapeutic index over the parent compound, EpoB, which is currently in clinical trials (5, 17). In the present study, we establish peloruside as a novel microtubule-stabilizing agent with potentially unique properties as compared with the other known microtubule-stabilizing drugs.

Peloruside A (peloruside; M_r 571.31) was isolated from the marine sponge M. hentscheli collected in Pelorus Sound off the northern coast of South Island, New Zealand. Peloruside was stored at −20°C as a 1 mm solution in absolute ethanol. Paclitaxel, purified tubulin, and mouse monoclonal antirat β-tubulin were purchased from Sigma Chemical Co. (St. Louis, MO). The NaBH₄ reduction product of peloruside was prepared as described previously (3). Less than 2% of the parent compound remained in the sample after the chemical reduction.

Lung adenocarcinoma H441 cells were grown as described previously (2, 3) on coverslips in 24-well plates and fixed with 0.5 ml of 3.7% formaldehyde in PBS for 45 min at 37°C. Cells were washed twice in PBS, permeabilized with T-PBS for 10 min, and washed twice again in PBS. The cells were then stained for β-tubulin or F-actin as follows.

For β-tubulin immunostaining, cells were incubated for 10 min at room temperature in 0.5 ml of 50% methanol, 1% H₂O₂ solution to remove endogenous horseradish peroxidase activity and then blocked in 0.2 ml of TSA blocking solution (Invitrogen, Auckland, New Zealand) per coverslip for 30 min at room temperature. Cells were then incubated with 0.2 ml of mouse anti-β-tubulin antibody (1:200 dilution in T-PBS; Sigma Chemical Co.) in a humidified incubator at 37°C for 1 h. Each coverslip was then incubated in 0.2 ml of secondary antibody solution (goat antimouse IgG-horseradish peroxidase; 1:200 dilution in T-PBS; Sigma Chemical Co.) containing 1 μl of RNase A (100 mg/ml) and 1 μl of PI (50 μg/ml) for 45 min at room temperature. Cells were then stained with FITC-tyramide solution according to the manufacturer’s instructions (Invitrogen). Coverslips were washed three times (5 min each) in 1 ml of PBS-T between each staining step.

For F-actin staining, 200 μl of Alexa 488 phalloidin (5 μl of 6.6 μm phalloidin stock solution in methanol added to 200 μl of PBS per coverslip; Molecular Probes, Eugene, OR) were added to cells, and cells were incubated at room temperature for 1 h in the dark. Five min before the end of the incubation period, 50 μl of PI solution (1 μg/ml in PBS) were added to each coverslip.

Stained cells on coverslips were mounted in PBS and visualized using a confocal microscope (TCS 4D; Leica Lasertechnik, Germany). Staining patterns were analyzed using NIH Image.

Using standard methodology, the DNA of cells was stained with PI, and the proportion of cells in different phases of the cell cycle was monitored by flow cytometry. Briefly, H441 cells were treated with 1 μm peloruside or 1 μm paclitaxel for 24 h. Adherent cells were collected by trypsinization and added to those collected from suspension. The cells were then fixed overnight with cold 70% ethanol and stained with PI solution consisting of 45 μg/ml PI, 10 μg/ml RNase A, and 0.1% glucose. After a 2-h incubation at room temperature, samples were analyzed in a FACSort flow cytometer (Becton Dickinson, Franklin Lakes, NJ).

A simple in situ cellular assay as described by Giannakakou et al. (13) was used, in which the shift in tubulin from depolymerized to polymerized forms was followed by electrophoresis and Western blotting of centrifuged particulate and cytosolic fractions. To summarize, 2 × 10⁶ untreated and drug-exposed human myeloid leukemic HL-60 cells cultured as described previously (2, 3) were lysed by exposure for 5 min at 37°C to 100 μl of hypotonic buffer [1 mm MgCl₂, 2 mm EGTA, 1% NP40, 2 mm phenylmethylsulfonyl fluoride, 1 μg/ml aprotinin, 2 μg/ml pepstatin, and 20 mm Tris-HCl (pH 6.8)]. The particulate fraction was separated from the soluble cytosolic fraction by high-speed centrifugation for 10 min in a benchtop centrifuge. Samples labeled “0 min” received drug immediately before collection of the cells. The processing of the cells to the critical centrifugation step required approximately 30 min. The pellet was dissolved in 100 μl of sample buffer (8 m urea, 4% 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonic acid, 3 m thiourea, and 40 mm DTT). Twenty μl of loading buffer were added to each 100-μl sample, and the samples were vortexed and then boiled for 5 min. Twenty μl of each sample were loaded onto an SDS/10% polyacrylamide gel and resolved by electrophoresis. β-Tubulin bands were identified by Western blotting using β-tubulin primary antibody (1:1000 dilution) following standard immunoblotting procedures with detection by enhanced chemiluminescence (Lumi-Light; Roche Diagnostics, Auckland, New Zealand).

Purified tubulin (approximately 7.5 mg of protein) containing approximately 15% microtubule-associated proteins was obtained from Sigma Chemical Co. and reconstituted in 0.1 m 4-morpholineethanesulfonic acid buffer (pH 6.8), 1 mm EGTA, 0.1 mm EDTA, 0.5 mm MgCl₂, 1 mm DTT, 0.1 mm GTP, 1 μg/ml leupeptin, 1 μg/ml aprotinin, and 100 μg/ml sucrose as stabilizer. The reconstituted tubulin was incubated at 37°C for 30 min in the presence of 10 μm peloruside or 10 μm paclitaxel. After a 30-min preincubation, one set of samples (untreated, peloruside-treated, and paclitaxel-treated samples) was removed for TEM. CaCl₂ (5 mm final concentration) was added to a second set, and the samples were incubated an additional 30 min at 37°C before preparation for TEM. A third set of preincubated samples was placed on ice and left for an additional 30 min at 0°C (no calcium added). For TEM, 2-μl aliquots were pipetted onto 400-mesh carbon- and Formvar-coated copper grids and left for 2 min at room temperature before blotting with filter paper. Each grid was stained with 5 μl of 1% uranyl acetate for 3 min at room temperature and then blotted with filter paper. Grids were air-dried overnight before examination in a Philips CM100 TEM operating at 80 kV.

The ring structures of peloruside A, paclitaxel, epothilone B, and laulimalide are compared in Fig. 1, along with the structure of the reduction product of peloruside in which the 6-membered pyranose ring was opened by chemical reduction, generating a secondary alcohol at C₉. The 48 h IC₅₀ values in HL-60 cells were estimated from MTT dye reduction assays as described previously (2, 3). The IC₅₀ values for peloruside, paclitaxel, and the reduction product of peloruside were 7 ± 4 (SE), 22 ± 8, and 221 ± 24 nm, respectively (n = 3 separate preparations, duplicate tests/preparation).

After a 2-day exposure to 100 nm peloruside or paclitaxel, multiple micronuclei were observed in H441 cells (Fig. 2, B and C). Longer exposures increased the number of micronuclei and the number of cells containing micronuclei. The center of the micronuclei stained strongly with PI, indicating that double-stranded DNA was present (data not shown). Peloruside (100 nm) caused the formation of microtubule fiber bundles in interphase H441 cells (Fig. 2,D) and multiple asters in mitotic cells (Fig. 2,E). In both peloruside- and paclitaxel-treated H441 cells, large intracellular rod-like fibers were observed by phase-contrast microscopy (Fig. 2, B and C). These cell-spanning rod-like fibers did not stain for either β-tubulin or F-actin and were also present in cells treated with 10 nm colchicine (data not shown).

Treatment of H441 cells with 1 μm peloruside or paclitaxel for 24 h led to partial cell cycle arrest in the G₂-M phase of the cell cycle (Fig. 3). The progression of cells into apoptosis/necrosis was seen as an increase in the number of cells in the subdiploid peak. The arrest in G₂-M was more complete for paclitaxel than for peloruside, and this difference was mirrored in the mitotic index of the cultures (Fig. 3). Cells treated with 1 μm peloruside had 34 ± 2% metaphase-arrested cells, whereas 64 ± 4% of cells showed metaphase arrest after paclitaxel treatment. Control H441 cultures without drug treatment typically had about 4% of cells in mitosis.

Tubulin in soluble and particulate fractions from HL-60 cells exposed to different concentrations of peloruside, paclitaxel, peloruside reduction product, or colchicine for 5 h were isolated and visualized by immunoblotting for β-tubulin (Fig. 4,A). Peloruside and paclitaxel caused similar, dose-dependent shifts of soluble tubulin to the particulate fraction. No detectable β-tubulin remained in the soluble fraction at 100 nm of either drug. The peloruside reduction product had no significant effect on the ratio of soluble to polymerized tubulin in HL-60 cells. Colchicine, as expected, caused depolymerization of tubulin, with most of the tubulin in the soluble fraction at a 1 μm concentration of drug. A 20-min time course was carried out in the presence of 1 μm peloruside and 1 μm paclitaxel (Fig. 4 B). By 5 min, both peloruside and paclitaxel had converted almost all detectable tubulin to the polymerized form.

Ten μm peloruside, like paclitaxel, caused purified tubulin to polymerize in solution into typical, long, straight microtubules at 37°C (Fig. 4 C). In the absence of drug, only a few sparse microtubules were seen by TEM. The microtubules formed in the presence of peloruside were stable at 0°C and in the presence of 5 mm CaCl₂ (data not shown).

In this study, we have shown that peloruside alters microtubule dynamics in a manner similar to that reported for paclitaxel by inducing tubulin polymerization in situ and in cell-free systems, causing cells to arrest in the G₂-M phase of the cell cycle. The identification of peloruside as a microtubule-stabilizing drug increases the small number of microtubule-stabilizing agents available for development as anticancer drugs. Despite the similarity of the primary mode of action of peloruside to the taxanes, epothilones, and laulimalides, it is interesting that in H441 cells, peloruside was less effective than paclitaxel at causing mitotic arrest. The unique structure and properties of peloruside may present novel benefits for anticancer targeting. Based on thin layer chromatography results in our laboratory, peloruside is less lipophilic than paclitaxel, and this property should aid the clinical application of peloruside or its analogues because some of the side effects of paclitaxel are related to its low aqueous solubility (6–11 μm; Ref. 10). Epothilones are reported to be 30–50 times more soluble than paclitaxel (5), and discodermolide is estimated to be 160-fold more soluble than paclitaxel, based on an indirect, fragment-based computational calculation (10). Laulimalides presumably have low aqueous solubility because they were selected in part on the basis of their lipophilic properties (4, 9).

Given the distinct structural differences between the various classes of microtubule-stabilizing drugs, structural comparisons between peloruside, epothilone B, and laulimalide should prove useful in deciphering the common pharmacophore (4, 5, 13, 15, 16). Peloruside, epothilones, and laulimalide are all polyoxygenated macrolides containing relatively hydrophobic side chains. The 6-membered pyranose ring of peloruside is essential for both its cytotoxicity (3) and microtubule-stabilizing effects (Fig. 4,A). Both peloruside A and the epothilones are C-15 (16-membered macrolides). When the structure of epothilone A is presented with the ketone and gemdimethyl in a similar conformation as peloruside A (Fig. 1), a similarity of the two structures is revealed. In particular, the pattern of oxygenation and substitution of C-3 to C-5 of epothilone and C-9 to C-11 of peloruside is the same. The side chains of the two molecules also coincide. The ester linkages, however, are in different positions and of opposite orientation, presumably indicating biosynthetic assembly in the opposite sense. Although peloruside appears to bind β-tubulin in the α,β-dimer in a manner similar to that of paclitaxel, further structure/function relationships are needed to model its active binding site.

In the present study, peloruside induced the formation of multiple micronuclei, multiple asters, microtubule bundles, rod-like fibers, and metaphase arrest in a manner similar to paclitaxel. Cell-type-specific differences exist in the reported responses to paclitaxel because some cells, such as HL-60 and the colon carcinoma cell line HT-29, arrest in metaphase and then undergo apoptosis, whereas other cells, such as K562 and the melanoma cell line SK-MEL-28, progress through metaphase and become polyploid in the presence of drug (18, 19). The apoptosis induced by peloruside (3) is presumably a consequence of either G₂-M block or the DNA damage due to abnormal mitotic arrest. Mitotic arrest often induces apoptosis in cultured cells (7, 20). With epothilone and paclitaxel, endonucleolytic cleavage of DNA, measured by the TUNEL assay, is only seen in G₂-M-blocked cells (7, 20). The apoptotic pathway for paclitaxel has been directly examined (reviewed in Ref. 20).

Our evidence that peloruside is a microtubule-stabilizing agent is based on an in situ cell assay (Fig. 4, A and B) and an in vitro polymerization assay (Fig. 4,C), in which a shift in tubulin from a soluble to a particulate form was observed. The conclusion that peloruside stabilizes microtubules in a manner similar to the taxanes and other microtubule-stabilizing drugs is also supported by the G₂-M cell cycle arrest data of Fig. 3. More direct measurements of peloruside-tubulin interactions in cell-free systems will be needed to fully describe the primary mode of action of peloruside, and these experiments are in progress.

We conclude that peloruside, a novel natural product with paclitaxel-like microtubule-stabilizing activity, represents a new drug in an elite group of drugs of major importance in the clinical treatment of solid tumors. Peloruside is structurally distinct and may present a unique profile of bioactivity that will add to that of the limited number of other known microtubule-stabilizing drugs available for pharmaceutical development. Important properties that need to be considered in further development of peloruside as an anticancer agent include the ease of producing synthetic analogues, its aqueous solubility, improved targeting of tumorigenic cells, reduced or unique susceptibility to the development of cellular resistance, and reduced toxicity to the body.